Note: Descriptions are shown in the official language in which they were submitted.
~tD~~~~3~3
_ 1 _
CASE 5933
NICKEL CATALYZED DISPLACEMENT REACTION
It has long been known that nickel catalyzes the
displacement of a butyl group as butene from a butyl
aluminum by ethylene (K. Ziegler, Brennstoff Chem., 35,
321 (1954)). The mechanism of this catalysis has been
extensively studied (K. Fischer et al. "Angewandte Chemie"
12 No. 12 pages 943-953 (December 1973)). It has been
proposed that the trialkyl aluminum reduces the nickel
compound to Ni metal which reacts with the a-olefin to
form an olefin complex which reacts with trialkyl aluminum
to displace the alkyls bonded to aluminum in an equilibri-
um reaction.
The nickel catalyzed alkyl displacement has not
achieved commercial significance because the catalyst is
extremely difficult to remove from the product and its
presence causes undesirable side reactions. These side
reactions are: reverse displacement, isomerization and
branched chain formation (Poe et al., Symposium on Produc-
tion and Use of Alpha Olefins, Div. of Petrol. Chem., Am.
Chem. Soc., Los Angeles, March 31 - April 5, 1963). Poe
et al. attempted to circumvent this problem by using a
heterogenous catalyst system with nickel deposited on a
support such as alumina. They reported high initial
displacement activity which sharply declined after 4-8
hours and could not be reactivated.
- 2 -
Davis et al., U. S. 3,391,219, resorted to non-
catalytic thermal displacement of alkyls of trialkyl
aluminum with a-olefins. Displacement temperatures in the
range of 280-320°C were required to effect the displace-
s ment followed by rapid cooling to about 120°C to minimize
isomerization and other side reactions of the effluent
a-olefin stream. Although very effective, this process
suffers from the high energy consumption required by the
high displacement temperature followed by the rapid
cooling.
From the above it can be seen that a need exists
for a process that can take advantage of the low tempera-
Lure displacement temperatures made possible by the use of
nickel catalysts while avoiding the undesirable side
reactions encountered in the past with the use of nickel
catalysts.
It has now been discovered that the nickel cata-
lyzed trialkyl aluminum displacement reaction can be
conducted without any substantial amount of side reactions
such as isomerization or dimerization by permitting the
displacement to proceed only until it is about at equili-
brium and then adding a catalyst poison in an amount
sufficient to deactivate the nickel catalyst.
A preferred embodiment of the invention is a
process for displacing alkyl groups from a first trialkyl
aluminum or mixture thereof, said process comprising
contacting said first trialkyl aluminum or mixture thereof
- 3 -
with an a-olefin or mixture thereof containing a different
distribution of carbon atoms than the alkyl groups in said
first trialkyl aluminum or mixture thereof, in the pres-
ence of a nickel catalyst at a temperature of -10 to 150°C
S whereby said a-olefins displace at least part of the alkyl
groups in said first trialkyl aluminum or mixture thereof
to form a second trialkyl aluminum or mixture thereof,
further characterized by adding a deactivating amount of a
catalyst poison selected from lead and compounds thereof
which are capable of deactivating said nickel catalyst
after the displacement reaction has substantially reached
equilibrium but before any significant isomerization of
the a-olefins to internal olefins has occurred.
The process is operable with any trialkyl aluminum
including those containing alkyl groups of 2 to 30 or more
carbon atoms including mixtures of different trialkyl
aluminum compounds and also trialkyl aluminum compounds
containing different alkyl groups in a single molecule.
Some examples are triethyl aluminum, tri-n-propyl alumi-
num, tri-isobutyl aluminum, tri-n-hexyl aluminum, tri-n-
octyl aluminum, tri-n-dodecyl aluminum, tri-n-tetradecyl
aluminum, tri-n-hexadecyl aluminum, tri-n-eicosyl alumi-
num, tri-n-docosanyl aluminum, tri-n-tricosyl aluminum,
diethyl n-butyl aluminum, ethyl butyl hexyl aluminum,
n-decyl n-dodecyl n-tetradecyl aluminum and the like
including mixtures thereof. Aluminum hydrides and alkyl
aluminum hydrides are known to inhibit the displacement
- 4 -
reaction so the trialkyl aluminum should not contain any
significant amount of hydride although it need not be
hydride-free.
In a preferred mode of operation the trialkyl
S aluminum is a mixture of trialkyl aluminums generated in
an ethylene chain growth process for making a--olefins as
described in Davis et al. U. S. 3,391,219. The displace-
ment reaction as used in that process results in peaking
of the more needed a-olefins. Since the trialkyl alumi-
nums in the process result from ethylene chain growth, the
alkyl groups will contain an even number of carbon atoms.
Triethyl aluminum in the effluent from the displacement
reaction is distilled and recycled to the feed for the
chain growth reaction. As a result, the resultant stream
will consist essentially of trialkyl aluminum compounds in
which the alkyl groups contain from 4-30 carbon atoms and.
have an even number of carbon atoms.
The a-olefin stream used in the displacement reac-
Lion can be any a-olefin stream such as those containing
2-30 carbon atoms such as ethylene, propylene, 1-butene,
1-hexene, 1-octene, 1-decene, 1-dodecene, 1-tetradecene,
1-hexadecene, 1-octadecene, 1-eicosene, 1-triacontene and
the like including mixtures thereof. The a-olefin stream
may include vinylidene, internal and tri-substituted
olefin diluents. When used as part of the chain growth
process as described in Davis et al. the a-olefin stream
will be primarily one or more a-olefins of even carbon
- 5 -
number from 4 to 30, more preferably from 4 to 16. In one
preferred embodiment the displacing olefin is ethylene.
In another highly preferred embodiment the displacing
olefin is 1-butene and the first trialkyl aluminum is a
tri-C2-30 alkyl aluminum. In still another preferred
embodiment the displacing olefin stream comprises mainly
1-butene, 1-hexene, 1-octene or 1-decene or mixtures of 2
or more of such olefins and the first trialkyl aluminum is
a tri-C4_16 alkyl aluminum. In still another preferred
embodiment the displacing olefin stream comprises mainly
1-dodecene, 1-tetradecene or 1-hexadecene or mixtures of 2
or more of such olefins and the first trialkyl aluminum is
a tri-C4_10 alkyl aluminum.
In order to achieve any useful result, the composi-
tion of the displacing olefin stream must be different
than the composition of the alkyl groups bonded to alumi-
num in the first trialkyl aluminum stream fed to the dis-
placement. If the molar amount of each individual a-
olefin in the olefin feed is the same as the equivalents
of each alkyl group in the first trialkyl aluminum, the
displacement product referred to as the second trialkyl
aluminum will be the same as the first trialkyl aluminum.
For example, if the a-olefin is 25 mole percent 1-butene,
mole percent 1-hexene and 50 mole percent 1-octene, and
25 the first trialkyl aluminum stream is a random mixture
containing 25 mole percent n-butyl groups, 25 mole percent
n-hexyl groups and 50 mole percent n-octyl groups, then
- 6 -
the product or second trialkyl aluminum will have about
the same composition, based on alkyl groups, as the first
trialkyl aluminum. This should be avoided since it does
not accomplish anything useful.
The amount of a-olefin should be in stoichiometric
excess over the amount required to replace all alkyl
groups. Preferably the amount of a-olefin should be at
least a 200 percent excess over the stoichiometric amount
required to replace all alkyl groups. Still more prefer-
ably the a-olefin feed should be at least a 500 percent
stoichiometric excess over the first trialkyl aluminum.
In this manner, since the displacement reaction is an
equilibrium reaction, the alkyl substitution in the second
trialkyl aluminum will more closely approach the distri-
bution of the a-olefin feed.
Any nickel catalyst can be used that has at least
some solubility in the displacement reaction system or
which reacts with the reaction mixture to form a nickel-
containing catalyst having at least some solubility.
Investigators in this area still differ as to what the
actual nickel catalyst is that exists in the system and
catalyzes the reaction. However, knowledge of what the
catalyst species is is not necessary to successfully carry
out the new process.
In a highly preferred embodiment, the nickel is
added in the form of a soluble nickel compound such as
nickel naphthenate, nickel bis(acetylacetonate),
nickelocene, bis(1,5-cyclooctadiene)nickel, nickel bis(N-
alkyl salicylaldimino) and the like.
The amount of nickel required to catalyze the reac
tion is very low, on the order of parts per million (ppm).
A useful range is 1-100 parts by weight nickel per million
parts of reaction mixture. A preferred range is 2-20 ppm.
A more preferred catalyst concentration is 2-l0 ppm.
The process is operable over a wide temperature
range. However, one of the main advantages of the nickel
catalyzed process over the non-catalyzed process is the
capability of operating at low to moderate temperatures.
For example Davis et al. disclose the non-catalytic dis-
placement reaction carried out at 280-320°C followed by a
rapid cool down to avoid side reactions. The nickel
catalyzed process proceeds at temperatures down to 15°C or
lower. There is no real upper limit as the reaction rate.
increases with temperature. As a practical matter, the
preferred temperature range is -10 to 150°C, more prefer-
ably 10-65°C and especially 20-35°C. This results in a
much more energy efficient process compared to the
non-catalytic process.
Both displacement and side reactions (e. g., isomeri-
zation, dimerization, chain growth) proceed concurrently.
However the displacement reaction rate is much higher than
the rate of the side reactions. This permits termination
of the displacement reaction after a time that allows it
to go substantially to completion and before a time in
'~'~~~i~3~3
- g _
which the side reactions, especially isomerization, become
significant. By "significant" is meant that the amount of
undesired by°products formed does not exceed an amount
which would render the olefin effluent stream unsuitable
for its intended purpose. In general the a-olefin product
should contain less than 25 weight percent newly formed
combined internal, tri-substituted and vinylidene olefins.
More preferably the a-olefin product should not exceed 20
weight percent newly formed combined internal, tri-substi-
tuted and vinylidene olefins. The preferred a-olefin
product is at least 80 weight percent and more preferably
85 weight percent vinyl a-olefin and more preferably at
least 90 weight percent vinyl a-olefin based on the tri-n-
alkylaluminum converted. The process is capable of making
a-olefin product that is at least 95 weight percent vinyl
a-olefin based on tri-n-alkylaluminum converted.
Since all rates vary with temperature and amount of
catalyst, the optimum time for termination under each
specific condition will require a minimal amount of experi-
mentation. In general when operating at 25°C, the reac-
tion should be terminated after a reaction period of about
seconds to 1 hour. A preferred reaction time is 1.75-
2.25 minutes. At higher 'temperatures, e.g., 50-100'C, the
preferred reaction time before side reactions become sig-
25 nificant will be shorter.
When the displacement has proceeded to the desired
extent, usually reaction equilibrium, a catalyst poison is
~~(D~~.~3~3
added in an amount that will deactivate the nickel cata-
lyst and prevent undesirable side reactions. These in-
elude lead and compounds thereof. Suitable lead compounds
are lead naphthenate, lead acetylacetonate, lead 2-ethyl-
hexanoate, and tetraethyl lead. Use of the metals as the
catalyst poison requires the metals to be in very finely
divided forms and requires a greater amount of the cata-
lyst poison. For example, amorphous lead metal was an
effective catalyst poison at a Pb/Ni atom ratio of about
500. The catalyst poisons which are effective at the
lowest concentrations have been lead compounds, e.g., lead
naphthenate, lead 2-ethylhexanoate and lead acetylaceto-
nate.
The amount of catalyst ~ioison should be an amount
that effectively inhibits all undesired side reactions.
With lead compounds a lead/nickel atom ratio of 1.0 has
been effective and even lower amounts may be effective.
Hence a useful Pb/Ni atom ratio is 0.5/1.0 to 5.0/1Ø
After the catalyst poison has been added, second
trialkyl aluminum product can be recovered by conventional
methods such as distillations. When lead compounds are
used as the poison, nickel and at least part of the lead
form a precipitate which can be removed by filtration.
The following examples show how the displacement
reaction is carried out and the results that can be
achieved.
i~~~~~~~
- 10 -
Examples 1-10
In a 6 dram vial fitted with a magnetic stirrer was
placed a mixture of 1-hexene (except where noted) and tri-
n-octyl aluminum (TNOA) at the mole ratio given in Table
I. To this was added a heptane solution of nickel (II)
naphthenate (except for Example 4 which used a heptane
solution of nickel bis(1,5-cyclooctadiene)) in an amount
to provide the amount of Ni in ppm shown in Table I. This
mixture was stirred at the temperature and time given in
Table T_. At that exact time a heptane solution of lead
(II) naphthenate was syringed into the reaction mixture in
Examples 2-3 and 5-10 to give the Pb/Ni atom ratio shown
in Table I. Lead (II) 2-ethylhexanoate was used in
Example 4. No lead poison was added to Example 1 which
serves as a comparative example.
Table
I
Exar~le 1-hexene/'fNOACatalystlTempC/ Pb/Ni
Atom
No. Mole Ratio Conc. Time Ratio
1 11.3 10.5 25/2 min none
2 11.3 10.5 25/30 sec 2.0
3 11.1 10.8 25/2 min 2.0
4 11.8 9.1 25/2 min 2.0
5 10.5 10.3 25/2 min 0.5
6 10.7 3.2 65/30 sec 2.0
7 10.7 3.3 65/30 sec 2.0
8 10.11 10.1 25/2 min 2.0
9 10.93 10.5 25/30 sec 2.0
10 4 9.9 100/2 min 2.0
- 11 -
1. Parts by weight Ni per million parts mixture.
2. Olefixi was 1-tetradecene.
3. Trialkyl aluminum was tri-n-butylaluminum.
4. Displacement olefin was ethylene at 225 psig in place of
1-hexene with 5.7 milimoles of TNOA.
The results for each example are given in Table II.
12 -
Table
II
Olefin
Co~ositiorrl~Area
Percent)
oR3Al At After
End Stirring
E~cam~le ConversionInitial of at Rx
Rx Temp
1 74 Vi 97.1 96.1 1.22
Int 1.4 2.3 95.6
Tri 0.2 0.2 1.7
Vd 1.3 1.4 1.4
2 43.23
3 72.4 Vi 97.7 - 97.44 97.35
Int 1.2 - 1.4 1.6
Tri 1.1 - 1.2 1.2
Vd <0.2 - <0.2 <0.2
4 75.8 Vi 96.7 96.2 95.94 96.36 96.17
Int 2.4 2.8 3.0 2.7 2.7
Tri <0.5 <0.5 <0.5 <0.5 <0.5
Vd 0.8 1.0 1.0 1.0 1.1
5 72.0 Vi 96.7 - 96.68 96.16 96.25
Int 2.4 - 2.4 2.6 2.5
Tri <0.5 - 0.5 0.8 0.9
Vd 0.8 - 0.5 0.5 0.4
6 73.6 Vi 96.8 - 95.59 95.210
Int 2.4 - 3.5 3.7
Tri <0.5 - <0.5 <0.5
Vd 0.8 - 1.0 1.1
7 70.0 Vi 98.0 - 9G.6~ 97.611
Int 0.7 - 1.4 1.4
Tri 0.7 - 1.1 0.6
Vd O.G - 0.8 0.4
8 74.5 Vi 86.1 - 85.71385.97
Int 5.5 - 6.1 6.0
Tri n.d.l2- n.d. n.d.
Vd 8.3 - 8.2 8.2
9 32.0 14
10 94.4 Vi - - 93.44 93.315
Int - - 4.2 4.3
'fri - - 0.9 0.9
Vd - - 1..5 1.5
~~~3~.~3~a
- 13 -
1. Vi = Vinyl olefin, Int. = Internal Olefin,
Tri = Tri-substituted olefin, Vd = Vinylidene
olefin.
2. 18 hours without Pb poison.
3. Conversion 5 min. after addition of Pb poison
- 44.2a, after
S 18 hr - 45.80, after 24 hours - 45.4%, after
42 hours -
45.7%.
4. 5 minutes after addition of Pb poison.
5. 18 hours after addition of Pb poison.
6. 1 hour after addition of Pb poison.
7. 72 hours after addition of Pb poison.
8. 4 minutes after addition of Pb poison.
9. 1 minute at 65C after addition of Pb poison.
10. 2 hours at 65C after addition of Pb poison.
11. Distilled at 25C/0.1 Torr.
1S 12. n.d. = not detected.
13. 1 minute after addition of Pb poison.
14. Conversion 5 minutes after addition of Pb poison
- 34%, after
10 minutes-35o, after 72 hours - 35%.
15. 4 days after addition of Pb poison.
Example 1 shows that at 25°C over a 2-minute period
1-hexene converts 74 percent of the TNOA forming an olefin
product that is (by gas chromatography) 96.1 area percent
vinyl olefin (i.e., R-CH=CH2), 2.3 area percent internal
olefin (i.e., R-CH=CH-R), 0.2 area percent tri-substituted
olefin (i.e., RRC=CHR) and 1.3 area percent vinylidene
olefin (i.e., R-C(=CH2)-R). However' on standing 18
hours with the nickel catalyst the vinyl content dropped
dramatically to only 1.2 area percent while internals
increased to 95.6 area percent. This is why the nickel
catalyzed displacement reaction has been of little use in
displacing alkyl groups from trialkyl aluminum (R3A1) to
form a different a-olefin. While the displacement goes
fast, the new a-olefin rapidly isomerizes to internal
olef ins.
- 14
Example 2 shows that after 30 seconds at 25°C, 43.2
percent of the initial TNOA is converted to a different
R3A1 by 1-hexene displacement. As shown in footnote 3,
this percent conversion remains almost constant through 42
hours showing that the Pb has deactivated the Ni catalyst.
Example 3 shows a 72.4 percent R3A1 conversion
after 2 minutes at 25°C to an initial olefin product that
is 97.7% vinyl, 1.2% internal, 1.1% tri-substituted and
<0.2% vinylidene. The significant feature here is that 5
minutes and 18 hours after the end of the displacement
reaction and injection of the Pb poison, the olefin olefin
composition remained almost constant. In fact vinyl
olefin content decreased only 0.4% over 18 hours at 25°C.
The remaining examples show the same high conver-
sion to a vinyl olefin-rich product that remains essential-
ly unisomerized over a long period due to the injection of
the Pb catalyst poison.
Example 10 is worth special notice for showing that
ethylene converts 94.4 percent of the TNOA after 2 minutes
at 100°C to a high a-olefin product which remains substan-
tially unchanged 4 days after the end o.f the displacement
and injection of the Pb solution.
The above results show that the new process makes
the known nickel catalyzed tria7.ky1 aluminum displacement
reaction extremely useful for alkyl displacement to pro-
duce a different a-olefin product that does not isomerize
or dimerize on standing.
